Showing 12 results for Classification
M. Hariri, S. B. Shokouhi, N. Mozayani,
Volume 4, Issue 3 (10-2008)
Abstract
Dealing with uncertainty is one of the most critical problems in complicated
pattern recognition subjects. In this paper, we modify the structure of a useful Unsupervised
Fuzzy Neural Network (UFNN) of Kwan and Cai, and compose a new FNN with 6 types of
fuzzy neurons and its associated self organizing supervised learning algorithm. This
improved five-layer feed forward Supervised Fuzzy Neural Network (SFNN) is used for
classification and identification of shifted and distorted training patterns. It is generally
useful for those flexible patterns which are not certainly identifiable upon their features. To
show the identification capability of our proposed network, we used fingerprint, as the most
flexible and varied pattern. After feature extraction of different shapes of fingerprints, the
pattern of these features, “feature-map”, is applied to the network. The network first
fuzzifies the pattern and then computes its similarities to all of the learned pattern classes.
The network eventually selects the learned pattern of highest similarity and returns its
specific class as a non fuzzy output. To test our FNN, we applied the standard (NIST
database) and our databases (with 176×224 dimensions). The feature-maps of these
fingerprints contain two types of minutiae and three types of singular points, each of them
is represented by 22×28 pixels, which is less than real size and suitable for real time
applications. The feature maps are applied to the FNN as training patterns. Upon its setting
parameters, the network discriminates 3 to 7 subclasses for each main classes assigned to
one of the subjects.
M. H. Sedaaghi,
Volume 5, Issue 1 (3-2009)
Abstract
Accurate gender classification is useful in speech and speaker recognition as
well as speech emotion classification, because a better performance has been reported when
separate acoustic models are employed for males and females. Gender classification is also
apparent in face recognition, video summarization, human-robot interaction, etc. Although
gender classification is rather mature in applications dealing with images, it is still in its
infancy in speech processing. Age classification, on the other hand, is also concerned as a
useful tool in different applications, like issuing different permission levels for different
aging groups. This paper concentrates on a comparative study of gender and age
classification algorithms applied to speech signal. Experimental results are reported for the
Danish Emotional Speech database (DES) and English Language Speech Database for
Speaker Recognition (ELSDSR). The Bayes classifier using sequential floating forward
selection (SFFS) for feature selection, probabilistic Neural Networks (PNNs), support
vector machines (SVMs), the K nearest neighbor (K-NN) and Gaussian mixture model
(GMM), as different classifiers, are empirically compared in order to determine the best
classifier for gender and age classification when speech signal is processed. It is proven that
gender classification can be performed with an accuracy of 95% approximately using
speech signal either from both genders or male and female separately. The accuracy for age
classification is about 88%.
A. Ghaffari, M. R. Homaeinezhad, M. Akraminia,
Volume 6, Issue 1 (3-2010)
Abstract
The aim of this study is to address a new feature extraction method in the area of the heart arrhythmia classification based on a metric with simple mathematical calculation called Curve-Length Method (CLM). In the presented method, curve length of the under study excerpted segment of signal is considered as an informative feature in which the effect of important geometric parameters of the original signal can be found. To show merits of the presented method, first the original electrocardiogram (ECG) in lead I is pre-processed by removing its baseline wander then by scaling it in the [-1,1] interval. In the next step, using a trous method, discrete wavelet scales 23 and 24 and smoothing function scale 22 are extracted. Afterwards, segments including samples of the QRS complex, P and T waves are estimated via an approximation criterion and CLM is implemented to extract corresponding features from aforementioned scales, smoothing function and also from each original segment. The resulted feature vector (including 12 components) is used to tune an Adaptive Network Fuzzy Inference System (ANFIS) classifier. The presented strategy is applied to classify four categories found in the MIT-BIH Arrhythmia Database namely as Atrial Premature Beat (APB), Left Bundle Branch Block (LBBB), Right Bundle Branch Block (RBBB) and Premature Ventricular Contraction (PVC) and average values of Se = 99.81%, P+ = 99.80%, Sp = 99.81% and Acc = 99.72% are obtained for sensitivity, positive predictivity, specifity and accuracy respectively showing marginal improvement of the heart arrhythmia classification performance.
M. R. Homaeinezhad, E. Tavakkoli, A. Afshar, A. Atyabi, A. Ghaffari,
Volume 7, Issue 2 (6-2011)
Abstract
The paper addresses a new QRS complex geometrical feature extraction technique as well as its application for electrocardiogram (ECG) supervised hybrid (fusion) beat-type classification. To this end, after detection and delineation of the major events of ECG signal via a robust algorithm, each QRS region and also its corresponding discrete wavelet transform (DWT) are supposed as virtual images and each of them is divided into eight polar sectors. Then, the curve length of each excerpted segment is calculated and is used as the element of the feature space. To increase the robustness of the proposed classification algorithm versus noise, artifacts and arrhythmic outliers, a fusion structure consisting of three Multi Layer Perceptron-Back Propagation (MLP-BP) neural networks with different topologies and one Adaptive Network Fuzzy Inference System (ANFIS) were designed and implemented. To show the merit of the new proposed algorithm, it was applied to all MIT-BIH Arrhythmia Database records and the discrimination power of the classifier in isolation of different beat types of each record was assessed and as the result, the average accuracy value Acc=98.27% was obtained. Also, the proposed method was applied to 8 number of arrhythmias (Normal, LBBB, RBBB, PVC, APB, VE, PB, VF) belonging to 19 number of the aforementioned database and the average value of Acc=98.08% was achieved. To evaluate performance quality of the new proposed hybrid learning machine, the obtained results were compared with similar peer-reviewed studies in this area.
M. Mollanezhad Heydar-Abadi , A. Akbari Foroud,
Volume 9, Issue 3 (9-2013)
Abstract
Fault classification in distance protection of transmission lines, with considering the wide variation in the fault operating conditions, has been very challenging task. This paper presents a probabilistic neural network (PNN) and new feature selection technique for fault classification in transmission lines. Initially, wavelet transform is used for feature extraction from half cycle of post-fault three phase currents at one end of line. In the proposed method three classifiers corresponding with three phases are used which fed by normalized particular features as wavelet energy ratio (WER) and ground index (GI). The PNNs are trained to provide faulted phase selection in different ten fault types. Finally, logic outputs of classifiers and GI identify the fault type. The feasibility of the proposed algorithm is tested on transmission line using PSCAD/EMTDC software. Variation of operating conditions in train cases is limited, but it is wide for test cases. Also, quantity of the test data sets is larger than the train data sets. The results indicate that the proposed technique is high speed, accurate and robust for a wide variation in operating conditions and noisy environments.
M. H Shakoor, F. Tajeripour,
Volume 11, Issue 3 (9-2015)
Abstract
In this paper, a special preprocessing operations (filter) is proposed to decrease
the effects of noise of textures. This filter using average of circular neighbor points (Cmean)
to reduce noise effect. Comparing this filter with other average filters such as square
mean filter and square median filter indicates that it provides more noise reduction and
increases the classification accuracy. After applying filter to noisy textures some Local
Binary Pattern (LBP) variants are used for feature extraction. The Implementation part for
noisy textures of Outex, UIUC and CUReT datasets shows that using proposed filter
increases the classification accuracy significantly. Furthermore, a simple and new technique
is proposed that increases the speed of c-mean filter noticeably.
A. A. Abedi, M. R. Mosavi, K. Mohammadi, M. R. Daliri,
Volume 12, Issue 3 (9-2016)
Abstract
One of the instruments for determination of position used in several applications is the Global Positioning System (GPS). With a cheap GPS receiver, we can easily find the approximate position of an object. Accuracy estimation depends on some parameters such as dilution of precision, atmospheric error, receiver noise, and multipath. In this study, position accuracy with GPS receiver is classified in three classes. Nine classification methods are utilized and compared. Finally, a new method is selected for classification. Results are verified with experimental data. Success rate for classificationis approximately 84%.
M. H. Refan, A. Dameshghi,
Volume 16, Issue 2 (6-2020)
Abstract
Geometric Dilution of Precision (GDOP) is a coefficient for constellations of Global Positioning System (GPS) satellites. These satellites are organized geometrically. Traditionally, GPS GDOP computation is based on the inversion matrix with complicated measurement equations. A new strategy for calculation of GPS GDOP is construction of time series problem; it employs machine learning and artificial intelligence methods for problem-solving. In this paper, the Time Delay Neural Network (TDNN) is introduced to the GPS satellite DOP classification. The TDNN has a memory for archiving past event that is critical in GDOP approximation. The TDNN approach is evaluated all subsets of satellites with the less computational burden. Therefore, the use of the inverse matrix method is not required. The proposed approach is conducted for approximation or classification of the GDOP. The experiments show that the approximate total RMS error of TDNN is less than 0.00022 and total performance of satellite classification is 99.48%.
M. Dodangeh, N. Ghaffarzadeh,
Volume 16, Issue 2 (6-2020)
Abstract
In this paper, a new fast and accurate method for fault detection, location, and classification on multi-terminal DC (MTDC) distribution networks connected to renewable energy and energy storages presented. MTDC networks develop due to some issues such as DC resources and loads expanding, and try to the power quality increasing. It is important to recognize the fault type and location in order to continue service and prevent further damages. In this method, a circuit kit is connected to the network. Fault detection is performed with the measurement of the current of the connected kits and the traveling-waves of the derivative of the fault current and applying to a mathematical morphology filter, in the Fault time. The type and location of faults determinate using circuit equations and current calculations. DC series and ground arc faults are considered as DC distribution network disturbances. The presented method was tested in an MTDC network with many faults. The results illustrate the validity of the proposed method. The main advantages of the proposed fault location and classification strategy are higher accuracy and speed than conventional approaches. This method robustly operates to changing in sampling frequency, fault resistance, and works very well in high impedance fault.
A. Saffari, S. H. Zahiri, M. Khishe,
Volume 18, Issue 1 (3-2022)
Abstract
In this paper, multilayer perceptron neural network (MLP-NN) training is used by the grasshopper optimization algorithm with the tuning of control parameters using a fuzzy system for the big data sonar classification problem. With proper tuning of these parameters, the two stages of exploration and exploitation are balanced, and the boundary between them is determined correctly. Therefore, the algorithm does not get stuck in the local optimization, and the degree of convergence increases. So the main aim is to get a set of real sonar data and then classify real sonar targets from unrealistic targets, including noise, clutter, and reverberation, using GOA-trained MLP-NN developed by the fuzzy system. To have accurate comparisons and prove the GOA performance developed with fuzzy logic (called FGOA), nine benchmark algorithms GOA, GA, PSO, GSA, GWO, BBO, PBIL, ES, ACO, and the standard backpropagation (BP) algorithm were used. The measured criteria are concurrency speed, ability to avoid local optimization, and accuracy. The results show that FGOA has the best performance for training datasets and generalized datasets with 96.43% and 92.03% accuracy, respectively.
M. Dodangeh, N. Ghaffarzadeh,
Volume 18, Issue 4 (12-2022)
Abstract
An intelligent strategy for the protection of AC microgrids is presented in this paper. This method was halving to an initial signal processing step and a machine learning-based forecasting step. The initial stage investigates currents and voltages with a window-based approach based on the dynamic decomposition method (DDM) and then involves the norms of the signals to the resultant DDM data. The results of the currents and voltages norms are applied as features for a topology data analysis algorithm for fault type classifying in the AC microgrid for fault location purposes. The Algorithm was tested on a microgrid that operates with precision equal to 100% in fault classification and a mean error lower than 20 m when forecasting the fault location. The proposed method robustly operates in sampling frequency, fault resistance variation, and noisy and high impedance fault conditions.
S. Prasad Tiwari,
Volume 19, Issue 3 (9-2023)
Abstract
In spite of the numerous benefits over the traditional power distribution system, protection of the microgrid is a challenging and complex task. The varying fault resistances due to dissimilar grounding conditions can affect the performance of the protection scheme. Under such conditions, the magnitude of the fault current can vary from lower to higher level. In addition to the above, the dissimilar magnitude of fault current during grid connected and islanded mode demands a protection scheme that can easily discriminate the mode of operation. The magnitude of fault current in grid-connected and islanded modes needs a robust protection scheme. In this regard, an ensemble of subspace kNN based robust protection scheme has been proposed to detect the faulty conditions of the microgrid. The tasks of the mode detection, fault detection/classification as well as faulty line identification has been carried out in the proposed work. In the proposed protection scheme, discrete wavelet transform (DWT) has been used for processing of the data. After recording the voltage and current signals at bus-1, the protection scheme has been validated. The validation of the protection scheme in Section 6 reveals that the protection scheme is efficiently working.
